{
  "name": "glintbase-scanner",
  "description": "Model Context Protocol server for the Glintbase Agent Readiness Scanner. Exposes the full scan pipeline — surface discovery, crawling, knowledge graph construction, agent journey simulation, ARS scoring, and remediation — as structured MCP tools. Runs locally via npx, no API keys required for core tools.",
  "version": "1.0.0",
  "package": "@glintbase/mcp",
  "repository": "https://github.com/glintbase/glintscanner/tree/main/mcp",
  "homepage": "https://glintbase.dev",
  "documentation": "https://glintbase.dev/docs",
  "transport": "stdio",
  "installation": {
    "mcpServers": {
      "glintbase": {
        "command": "npx",
        "args": ["-y", "@glintbase/mcp"]
      }
    }
  },
  "tools": [
    {
      "name": "discover_surfaces",
      "description": "Discover all machine-readable entrypoints for a product's developer ecosystem: llms.txt, OpenAPI specs, MCP configs, sitemaps, docs roots, GitHub repos, SDK packages, auth flows, changelogs, and more."
    },
    {
      "name": "check_reachability",
      "description": "Quick single-URL reachability check for AI agents. Tests accessibility, measures latency, detects soft-404s, and classifies content type."
    },
    {
      "name": "parse_spec",
      "description": "Parse and analyze a machine-readable specification file. Supports OpenAPI, llms.txt, and MCP configs."
    },
    {
      "name": "crawl_pages",
      "description": "Crawl the product's documentation ecosystem with a priority-queue budget. Extracts titles, content quality metrics, internal links, and machine-readability scores."
    },
    {
      "name": "deep_crawl",
      "description": "In-depth crawl that recovers real page content from JS-rendered documentation sites (Next.js, Docusaurus, Nextra, SPA shells) via embedded data extraction."
    },
    {
      "name": "build_knowledge_graph",
      "description": "Build a semantic knowledge graph from crawled pages. Maps concepts, API operations, and topics; identifies isolated clusters and coverage gaps."
    },
    {
      "name": "run_journeys",
      "description": "Run deterministic agent journey simulations against the knowledge graph. Tests whether an AI agent can complete real integration tasks using only the available documentation."
    },
    {
      "name": "score_readiness",
      "description": "Calculate the composite Agent Readiness Score (ARS 1.0): weighted 0-100 score, band classification, and per-dimension breakdown."
    },
    {
      "name": "get_remediation",
      "description": "Get prioritized, actionable remediation advice for improving agent readiness, with estimated effort and expected score impact per fix."
    }
  ]
}